gboolean -> bool conversion
Hi,
I'd like to convert a bunch of gboolean decleration to stdbool in our ConnMan code base. I am a noob with coccinelle. I have partial success so far but I do not thing I am doing the right thing. So any advice is welcome :)
This here is my current coccinelle script for converting gboolean used as stack variable. I have also one for structs, which gave my good results. Another one is needed then for the function arguments. But let's first have a look on this part.
@r1@
position p;
typedef gboolean;
identifier func,x;
@@
func(...) {
<...
gboolean@p x;
...>
}
@r2@
position r1.p;
typedef bool;
@@
- gboolean at p
+ bool
@r3@
identifier r1.x;
@@
(
- x = FALSE
+ x = false
|
- x = TRUE
+ x = true
)
int main(int argc, char *argv[])
{
gboolean b = TRUE;
return 0;
}results in
int main(int argc, char *argv[])
{
- gboolean b = TRUE;
+ bool b = true;
return 0; }
which is what I wanted. Now the problem is that I really have no clue to mach on things like
int main(int argc, char *argv[])
{
gboolean b = TRUE, c = FALSE;
return 0;
}Any idea?
cheers,
daniel
Quoted 61 lines · Daniel Wagner
Hi,
I'd like to convert a bunch of gboolean decleration to stdbool in our ConnMan code base. I am a noob with coccinelle. I have partial success so far but I do not thing I am doing the right thing. So any advice is welcome :)
This here is my current coccinelle script for converting gboolean used as stack variable. I have also one for structs, which gave my good results. Another one is needed then for the function arguments. But let's first have a look on this part.
@r1@
position p;
typedef gboolean;
identifier func,x;
@@
func(...) {
<...gboolean at p x;...>
}@r2@
position r1.p;
typedef bool;
@@
- gboolean at p
+ bool@r3@
identifier r1.x;
@@
(
- x = FALSE
+ x = false
|
- x = TRUE
+ x = true
)int
main(int argc, char *argv[]){ gboolean b = TRUE; return 0; }results in
int
main(int argc, char *argv[])
{(quoted diff) +1
-1- gboolean b = TRUE; + bool b = true; return 0;}
which is what I wanted. Now the problem is that I really have no clue to mach on things like
Quoted 8 lines
int
main(int argc, char *argv[]){ gboolean b = TRUE, c = FALSE; return 0; }Any idea?
Did you try just
@@
@@
- gboolean
+ bool
Also, for yoru TRUE -> true rule, you could consider
@@
gboolean x;
@@
x =
- TRUE
+ true
That is, you don't have to find the type declaration and then the identifier. You can describe the type of an arbitrary expression x.
julia
Hi Julia,
Quoted 70 lines · Julia Lawall
Quoted 59 lines
I'd like to convert a bunch of gboolean decleration to stdbool in our ConnMan code base. I am a noob with coccinelle. I have partial success so far but I do not thing I am doing the right thing. So any advice is welcome :)
This here is my current coccinelle script for converting gboolean used as stack variable. I have also one for structs, which gave my good results. Another one is needed then for the function arguments. But let's first have a look on this part.
@r1@
position p;
typedef gboolean;
identifier func,x;
@@
func(...) {
<...gboolean at p x;...>
}@r2@
position r1.p;
typedef bool;
@@
- gboolean at p
+ bool@r3@
identifier r1.x;
@@
(
- x = FALSE
+ x = false
|
- x = TRUE
+ x = true
)int
main(int argc, char *argv[]){ gboolean b = TRUE; return 0; }results in
int
main(int argc, char *argv[])
{(quoted diff) +1
-1- gboolean b = TRUE; + bool b = true; return 0;}
which is what I wanted. Now the problem is that I really have no clue to mach on things like
Quoted 8 lines
int
main(int argc, char *argv[]){ gboolean b = TRUE, c = FALSE; return 0; }Any idea?
Did you try just
Quoted 7 lines
@@
@@- gboolean
+ boolAlso, for yoru TRUE -> true rule, you could consider
Yes, the problem is that I can't replace all gbooleans because we are GLib, e.g. we are using quite a few timers
guint g_timeout_add (guint interval, GSourceFunc function,
gpointer data);
gboolean (*GSourceFunc) (gpointer user_data);
https://developer.gnome.org/glib/2.36/glib-The-Main-Event-Loop.html#g-timeout-add
So these gboolean should not be changed. BTW, when I tried your probosal it would not change this here
gboolean b,c;
changing the rules to
- gboolean ++ bool
resulted int
bool bool b,c;
Also not what I wanted :)
Anyway, the main problem I think is that I do not want to change all gboolean, only the ones in structs and the ones declared on the stack. Maybe a few in fucntion signatures but that I can do by hand. I just dont want to change 1000 lines by hand, that is far to dangerous :)
Quoted 10 lines
@@
gboolean x;
@@x =
- TRUE
+ trueThat is, you don't have to find the type declaration and then the identifier. You can describe the type of an arbitrary expression x.
Ah okay. I wanted to make sure I only change those assignments which did change.
cheers,
daniel
Quoted 103 lines · Daniel Wagner
Hi Julia,
Quoted 74 lines · Julia Lawall
Quoted 63 lines
I'd like to convert a bunch of gboolean decleration to stdbool in our ConnMan
code base. I am a noob with coccinelle. I have partial success so far but I do
not thing I am doing the right thing. So any advice is welcome :)This here is my current coccinelle script for converting gboolean used as stack variable. I have also one for structs, which gave my good results. Another one is needed then for the function arguments. But let's first have a
look on this part.@r1@
position p;
typedef gboolean;
identifier func,x;
@@
func(...) {
<...gboolean at p x;...>
}@r2@
position r1.p;
typedef bool;
@@
- gboolean at p
+ bool@r3@
identifier r1.x;
@@
(
- x = FALSE
+ x = false
|
- x = TRUE
+ x = true
)int
main(int argc, char *argv[]){ gboolean b = TRUE; return 0; }results in
int
main(int argc, char *argv[])
{(quoted diff) +1
-1- gboolean b = TRUE; + bool b = true; return 0;}
which is what I wanted. Now the problem is that I really have no clue to mach
on things likeQuoted 8 lines
int
main(int argc, char *argv[]){ gboolean b = TRUE, c = FALSE; return 0; }Any idea?
Did you try just
Quoted 7 lines
@@
@@- gboolean
+ boolAlso, for yoru TRUE -> true rule, you could consider
Yes, the problem is that I can't replace all gbooleans because we are GLib, e.g. we are using quite a few timers
guint
g_timeout_add(guint interval, GSourceFunc function,gpointer data);gboolean (*GSourceFunc) (gpointer
user_data);https://developer.gnome.org/glib/2.36/glib-The-Main-Event-Loop.html#g-timeout-add
So these gboolean should not be changed. BTW, when I tried your probosal it would not change this here
gboolean b,c;changing the rules to
Quoted 8 lines
- gboolean ++ boolresulted int
bool bool b,c;Also not what I wanted :)
OK, I will get back to you.
julia
Just for the record, I played a bit more around.
@@
@@
- gboolean
+ bool
@@
gboolean x;
@@
x =
(
- TRUE
+ true
|
- FALSE
+ false
)
applied on
static gboolean is_enabled(gboolean var)
{
return var;
}int main(int argc, char *argv[])
{
gboolean b = TRUE;
gboolean c, d;
gboolean e = FALSE, g, f = TRUE;
gboolean h = is_ensbled(b);
return 0;
}results in
+++ /tmp/cocci-output-20339-560145-test.c
@@ -5,9 +5,9 @@ static gboolean is_enabled(gboolean var)
int main(int argc, char *argv[])
{
- gboolean b = TRUE;
+ gboolean b = true;
gboolean c, d;
- gboolean e = FALSE, g, f = TRUE;
+ gboolean e = false, g, f = true;
gboolean h = is_ensbled(b);
return 0;How is the following:
@@
gboolean x;
@@
x =
(
- TRUE
+ true
|
- FALSE
+ false
)
@@
typedef bool;
@@
- gboolean
+ bool
Note that the order of the rules is important. In your rules, you converted gboolean to bool first, so the second rule would not find any gbooleans.
Another problem with your rules was that it thought that gboolean and bool were expressions. So the first rule needed typedef declaration. In my rule, I use gboolean as the type of a metavariable in the first rule, so Coccinelle considers it to be a type thereafter. But I need the typedef on bool in the second rule.
Is that good enough for your purpose, or are too many values and declarations being converted?
julia
Quoted 15 lines · Julia Lawall
How is the following:
@@
gboolean x;
@@x =
(
- TRUE
+ true
|
- FALSE
+ false
)
Quoted 8 lines
@@
typedef bool;
@@- gboolean
+ boolNote that the order of the rules is important.
Yeah, I should have noted that.
Quoted 3 lines
In your rules, you
converted gboolean to bool first, so the second rule would not find any gbooleans.
Yep, that works better, though now I hit the 'already tagged' problem. Changing the second rule to
-gboolean
++ bool
Let's coccinelle run but the result is not perfect.
Quoted 5 lines
Another problem with your rules was that it thought that gboolean and bool were expressions. So the first rule needed typedef declaration. In my rule, I use gboolean as the type of a metavariable in the first rule, so Coccinelle considers it to be a type thereafter. But I need the typedef on bool in the second rule.
Thanks for the explanation. It really helps!
Quoted 2 lines
Is that good enough for your purpose, or are too many values and declarations being converted?
I have updated my sample program, which contains all our usage pattern of gboolean.
struct info {
gboolean data;
};static gboolean timer_cb(gpointer user_data)
{
struct info *info = user_data;
info->data = FALSE;
return FALSE;
}static gboolean is_enabled(gboolean flag)
{
return flag != TRUE;
}int main(int argc, char *argv[])
{
gboolean b = TRUE;
gboolean c, d;
gboolean e = FALSE, g, f = TRUE;
gboolean h = is_ensbled(b);
struct info *info;
info = g_new0(struct info, 1);
info->data = TRUE;
g_timeout(delay, timer_cb, &info);
return 0;
}The resulting patch (with the ++bool version) is:
tmp/cocci-output-21088-1cd7bc-test.c renamed from test.c +9 -9
@@ -1,31 +1,31 @@struct info { -gboolean data; +bool data; }; -static gboolean timer_cb (gpointer user_data) +static bool timer_cb (gpointer user_data) {struct info *info = user_data; - info->data = FALSE; + info->data = false;return FALSE; } -static gboolean is_enabled (gboolean flag) +static bool is_enabled (bool flag) {return !flag; }int main (int argc,char *argv[]) { -gboolean b = TRUE; -gboolean c, d; -gboolean e = FALSE, g, f = TRUE; -gboolean h =is_ensbled (b); +bool b = true; + bool bool c, d; + bool bool bool e = false, g, f = true; +bool h =is_ensbled (b);struct info *info; info = g_new0(struct info, 1); - info->data = TRUE; + info->data = true;g_timeout (delay, timer_cb, &info);
Obviously, the 'bool bool bool' stuff is not correct, but I can catch that via the compiler.
timer_cb() should not be changed. Is there maybe a way of blacklisting certain matches? I see there are some 'when' examples but I do not really growk them right now. Sorry for beeing so noobish...
cheers,
daniel
Maybe the patch below will get rid of the already tagged token problem and allow you to get rid of the ++..
julia
engine/transformation_c.ml +22 -0
@@ -340,6 +340,28 @@ module XTRANS = struct Some(Ast_cocci.CONTEXT(old_pos,old_modif), (clean_env tin.binding)::oldenvs) + (* non ++ but same modif for two reasons *) + | (Ast_cocci.MINUS(old_pos,old_inst,old_adj,old_modif), + Ast_cocci.MINUS(new_pos,new_inst,new_adj,new_modif)) + when old_pos = new_pos && + old_modif = strip_minus_code new_modif && + List.mem (clean_env tin.binding) oldenvs -> + + cocciinforef := + Some(Ast_cocci.MINUS(old_pos,Common.union_set old_inst new_inst, + old_adj,old_modif), + oldenvs) + + | (Ast_cocci.CONTEXT(old_pos,old_modif), + Ast_cocci.CONTEXT(new_pos,new_modif)) + when old_pos = new_pos && + old_modif = strip_context_code new_modif && + List.mem (clean_env tin.binding) oldenvs -> + (* iteration only allowed on context; no way to replace something + more than once; now no need for iterable; just check a flag *) + + cocciinforef := Some(Ast_cocci.CONTEXT(old_pos,old_modif),oldenvs) + | _ -> (* coccionly: if !Flag.sgrep_mode2
Hi Julia,
Quoted 2 lines · Julia Lawall
Maybe the patch below will get rid of the already tagged token problem and allow you to get rid of the ++..
Thanks for the quick fix. Indeed, this fixes the 'already tagged' problem. The result is now:
struct info {
- gboolean data;
+ bool data;
};
-static gboolean timer_cb(gpointer user_data)
+static bool timer_cb(gpointer user_data)
{
struct info *info = user_data;- info->data = FALSE;
+ info->data = false;
return FALSE;
}
-static gboolean is_enabled(gboolean flag)
+static bool is_enabled(bool flag)
{
return !flag;}
int main(int argc, char *argv[])
{
- gboolean b = TRUE;
- gboolean c, d;
- gboolean e = FALSE, g, f = TRUE;
- gboolean h = is_ensbled(b);
+ bool b = true;
+ bool c, d;
+ bool e = false, g, f = true;
+ bool h = is_ensbled(b);
struct info *info;
info = g_new0(struct info, 1);- info->data = TRUE;
+ info->data = true;
g_timeout(delay, timer_cb, &info);
I am pondering if it would be possible to say
- gboolean
when != 'static gboolean func(gpointer)'
+ bool
cheers,
daniel
ps: I the install step didn't work for me:
$ make install cp: cannot create regular file ?ocaml/coccilib/coccilib.cmi?: No such file or directory make: *** [ocaml/coccilib/coccilib.cmi] Error 1
Doing a mkdir before the install step fixed it for me.
$ mkdir ocaml/coccili $ make install
Quoted 5 lines
I am pondering if it would be possible to say
- gboolean
when != 'static gboolean func(gpointer)'+ bool
No, this is not possible. Is the idea that you don't want to change the return type of a static function? Then you can record the position of the gbooleans that you don't want to change, and only change the others.
An example is as follows:
@keep@
identifier f;
position p;
typedef gpointer;
identifier i;
@@
static gboolean at p f(gpointer i) { ... }
@@
typedef bool;
position p != keep.p;
@@
- gboolean at p
+ bool
Note that you could never refer to
static gboolean func(gpointer)
because it is not a complete syntactic unit. If you wanted to mention a prototype, you would need a ; after it. If you wanted to mention the definition, then you need to put the complete definition, as shown above.
julia
Hi Julia,
Quoted 9 lines · Julia Lawall
Quoted 5 lines
I am pondering if it would be possible to say
- gboolean
when != 'static gboolean func(gpointer)'+ bool
No, this is not possible. Is the idea that you don't want to change the return type of a static function? Then you can record the position of the gbooleans that you don't want to change, and only change the others.
It is so obvious when you explain it. Thanks a lot.
Quoted 11 lines
An example is as follows:
@keep@
identifier f;
position p;
typedef gpointer;
identifier i;
@@static gboolean at p f(gpointer i) { ... }
Quoted 8 lines
@@
typedef bool;
position p != keep.p;
@@- gboolean at p
+ bool
With some small modification this works fine now.
The only outstanding problem I'd like to address is patching the function calls.
Let's say I have this deceleration (which will be changed by the above rules):
int foo(gboolean b);
and some place I call then
foo(FALSE);
This should be changed to
foo(false);
Quoted 7 lines
Note that you could never refer to
static gboolean func(gpointer)
because it is not a complete syntactic unit. If you wanted to mention a prototype, you would need a ; after it. If you wanted to mention the definition, then you need to put the complete definition, as shown above.
I see. Thanks again for explaining.
thanks
daniel
Quoted 55 lines · Daniel Wagner
Hi Julia,
Quoted 9 lines · Julia Lawall
Quoted 5 lines
I am pondering if it would be possible to say
- gboolean
when != 'static gboolean func(gpointer)'+ bool
No, this is not possible. Is the idea that you don't want to change the return type of a static function? Then you can record the position of the gbooleans that you don't want to change, and only change the others.
It is so obvious when you explain it. Thanks a lot.
Quoted 11 lines
An example is as follows:
@keep@
identifier f;
position p;
typedef gpointer;
identifier i;
@@static gboolean at p f(gpointer i) { ... }
Quoted 8 lines
@@
typedef bool;
position p != keep.p;
@@- gboolean at p
+ boolWith some small modification this works fine now.
The only outstanding problem I'd like to address is patching the function calls.
Let's say I have this deceleration (which will be changed by the above rules):
int foo(gboolean b);and some place I call then
foo(FALSE);This should be changed to
foo(false);Quoted 2 lines
Note that you could never refer to
Quoted 7 lines
Quoted 5 lines
static gboolean func(gpointer)
because it is not a complete syntactic unit. If you wanted to mention a prototype, you would need a ; after it. If you wanted to mention the definition, then you need to put the complete definition, as shown above.
I see. Thanks again for explaining.
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...) { ... }
@@
identifier r.f;
expression list [r.n] es;
@@
f(es,
(
- FALSE
+ false
|
- TRUE
+ true
)
,...)
I have not tested this, but it should work. Write back if you have problems.
julia
Quoted 2 lines · Julia Lawall
I have not tested this, but it should work. Write back if you have problems.
Sure, will do. Though today I don't find time to test.
thanks,
daniel
Hi Julia,
Quoted 24 lines · Julia Lawall
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...){ ... }@@
identifier r.f;
expression list [r.n] es;
@@f(es,
(
- FALSE
+ false
|
- TRUE
+ true
)
,...)
Quoted 2 lines
I have not tested this, but it should work. Write back if you have problems.
This rule is a half success. The main problem stems from the fact, that a function is declared in a header, e.g.
void connman_foo(gboolean foo);
and the caller just sees that and not the definition. If I understood this correctly the 'f(ps, gboolean i, ...) { ... }' will not match.
Another boolean transformation I do is with connman_bool_t to bool. Coccinelle is kind of unhappy with this function:
https://git.kernel.org/cgit/network/connman/connman.git/tree/vpn/vpn-agent.c#n38
parse error
= error in vpn/vpn-agent.c; set verbose_parsing for more info
badcount: 20
bad: #include "vpn.h"
bad:
bad: connman_bool_t vpn_agent_check_reply_has_dict(DBusMessage *reply)
bad: {
bad: const char *signature = DBUS_TYPE_ARRAY_AS_STRING
BAD:!!!!! DBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRING
bad: DBUS_TYPE_STRING_AS_STRING
bad: DBUS_TYPE_VARIANT_AS_STRING
bad: DBUS_DICT_ENTRY_END_CHAR_AS_STRING;
bad:
bad: if (dbus_message_has_signature(reply, signature) == TRUE)
bad: return TRUE;
bad:
bad: connman_warn("Reply %s to %s from %s has wrong signature %s",
bad: signature,
bad: dbus_message_get_interface(reply),
bad: dbus_message_get_sender(reply),
bad: dbus_message_get_signature(reply));bad:
bad: return FALSE;
bad: }
Is there a way to make the parser happy?
cheers,
daniel
Quoted 30 lines · Daniel Wagner
Hi Julia,
Quoted 24 lines · Julia Lawall
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...){ ... }@@
identifier r.f;
expression list [r.n] es;
@@f(es,
(
- FALSE
+ false
|
- TRUE
+ true
)
,...)Quoted 2 lines
I have not tested this, but it should work. Write back if you have problems.
Quoted 7 lines
This rule is a half success. The main problem stems from the fact, that a function is declared in a header, e.g.
void connman_foo(gboolean foo);and the caller just sees that and not the definition. If I understood this correctly the '
f(ps, gboolean i, ...){ ... }' will not match.
OK, you can make a rule that matches a prototype as well.
Quoted 26 lines
Another boolean transformation I do is with
connman_bool_tto bool. Coccinelle is kind of unhappy with this function:https://git.kernel.org/cgit/network/connman/connman.git/tree/vpn/vpn-agent.c#n38
parse error
= error invpn/vpn-agent.c; setverbose_parsingfor more info
badcount: 20
bad: #include "vpn.h"
bad:
bad:connman_bool_tvpn_agent_check_reply_has_dict(DBusMessage *reply)
bad: {
bad: const char *signature =DBUS_TYPE_ARRAY_AS_STRING
BAD:!!!!!DBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRINGbad: DBUS_TYPE_STRING_AS_STRING bad: DBUS_TYPE_VARIANT_AS_STRING bad: DBUS_DICT_ENTRY_END_CHAR_AS_STRING; bad: bad: if (dbus_message_has_signature(reply, signature) == TRUE) bad: return TRUE; bad: bad: connman_warn("Reply %s to %s from %s has wrong signature %s", bad: signature, bad: dbus_message_get_interface(reply), bad: dbus_message_get_sender(reply), bad: dbus_message_get_signature(reply));
Quoted 5 lines
bad:
bad: return FALSE;
bad: }Is there a way to make the parser happy?
Yes, you need to give some macro definitions. Make a copy of standard.h, add what you need and give the name of the file with the option --macro-file-builtins.
Concretely, you want to put the following definitions:
#define DBUS_TYPE_ARRAY_AS_STRING 0
#define DBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRING
#define DBUS_TYPE_STRING_AS_STRING
#define DBUS_TYPE_VARIANT_AS_STRING
#define DBUS_DICT_ENTRY_END_CHAR_AS_STRING
julia
Hi Julia,
Quoted 35 lines · Julia Lawall
Quoted 33 lines · Daniel Wagner
Hi Julia,
Quoted 24 lines · Julia Lawall
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...){ ... }@@
identifier r.f;
expression list [r.n] es;
@@f(es,
(
- FALSE
+ false
|
- TRUE
+ true
)
,...)Quoted 2 lines
I have not tested this, but it should work. Write back if you have problems.
This rule is a half success. The main problem stems from the fact, that a function is declared in a header, e.g.
Quoted 6 lines
Quoted 4 lines
void connman_foo(gboolean foo);and the caller just sees that and not the definition. If I understood this correctly the '
f(ps, gboolean i, ...){ ... }' will not match.OK, you can make a rule that matches a prototype as well.
I am still struggling with creating a rule for a prototype. Rereading examples and documentation.
So I am only allowed to write a prototype (funproto, right?) in the declaration section (common_decl)? How do I write the rule?
(I am really bad not getting this done by myself.)
Quoted 36 lines
Quoted 26 lines
Another boolean transformation I do is with
connman_bool_tto bool. Coccinelle is kind of unhappy with this function:https://git.kernel.org/cgit/network/connman/connman.git/tree/vpn/vpn-agent.c#n38
parse error
= error invpn/vpn-agent.c; setverbose_parsingfor more info
badcount: 20
bad: #include "vpn.h"
bad:
bad:connman_bool_tvpn_agent_check_reply_has_dict(DBusMessage *reply)
bad: {
bad: const char *signature =DBUS_TYPE_ARRAY_AS_STRING
BAD:!!!!!DBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRINGbad: DBUS_TYPE_STRING_AS_STRING bad: DBUS_TYPE_VARIANT_AS_STRING bad: DBUS_DICT_ENTRY_END_CHAR_AS_STRING; bad: bad: if (dbus_message_has_signature(reply, signature) == TRUE) bad: return TRUE; bad: bad: connman_warn("Reply %s to %s from %s has wrong signature %s", bad: signature, bad: dbus_message_get_interface(reply), bad: dbus_message_get_sender(reply), bad: dbus_message_get_signature(reply));Quoted 5 lines
bad:
bad: return FALSE;
bad: }Is there a way to make the parser happy?
Yes, you need to give some macro definitions. Make a copy of
standard.h, add what you need and give the name of the file with the option --macro-file-builtins.
Quoted 7 lines
Concretely, you want to put the following definitions:
#define
DBUS_TYPE_ARRAY_AS_STRING0
#defineDBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRING
#defineDBUS_TYPE_STRING_AS_STRING
#defineDBUS_TYPE_VARIANT_AS_STRING
#defineDBUS_DICT_ENTRY_END_CHAR_AS_STRING
That would work in the example but it would not do it for
https://git.kernel.org/cgit/network/connman/connman.git/tree/plugins/vpn.c#n666
const char *signature = DBUS_TYPE_ARRAY_AS_STRING DBUS_STRUCT_BEGIN_CHAR_AS_STRING DBUS_TYPE_OBJECT_PATH_AS_STRING DBUS_TYPE_ARRAY_AS_STRING DBUS_DICT_ENTRY_BEGIN_CHAR_AS_STRING DBUS_TYPE_STRING_AS_STRING DBUS_TYPE_VARIANT_AS_STRING DBUS_DICT_ENTRY_END_CHAR_AS_STRING DBUS_STRUCT_END_CHAR_AS_STRING;
So my solution was to pull in the dbus-protocol.h header which defines them as strings and everyting works fine.
cheers,
daniel
Quoted 7 lines
I am still struggling with creating a rule for a prototype. Rereading examples and documentation.
So I am only allowed to write a prototype (funproto, right?) in the declaration section (
common_decl)? How do I write the rule?(I am really bad not getting this done by myself.)
Could you send your attempt? It should be possible to write a function prototype pattern just like any other pattern, but it is true that I haven't done that very much, so it might be deficient in some way.
julia
Quoted 11 lines · Julia Lawall
Quoted 7 lines
I am still struggling with creating a rule for a prototype. Rereading examples and documentation.
So I am only allowed to write a prototype (funproto, right?) in the declaration section (
common_decl)? How do I write the rule?(I am really bad not getting this done by myself.)
Could you send your attempt? It should be possible to write a function prototype pattern just like any other pattern, but it is true that I haven't done that very much, so it might be deficient in some way.
So my approach was to try this here:
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...);
And after going over the documentation I think I see my error though I have not figured yet how to do it right.
cheers,
daniel
Quoted 17 lines · Daniel Wagner
Quoted 4 lines · Julia Lawall
Quoted 4 lines
I am still struggling with creating a rule for a prototype. Rereading examples
and documentation.Quoted 8 lines
Quoted 4 lines
So I am only allowed to write a prototype (funproto, right?) in the declaration section (
common_decl)? How do I write the rule?(I am really bad not getting this done by myself.)
Could you send your attempt? It should be possible to write a function prototype pattern just like any other pattern, but it is true that I haven't done that very much, so it might be deficient in some way.
So my approach was to try this here:
Quoted 7 lines
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...);
Just add the metavariable declaration
type T;
and T to the left of f in the prototype. For function definitions, it is OK with the return type being absent, but apparently not for prototypes. Leaving it out probably causes conflicts for the parser, because a prototype looks like a function call.
julia
Quoted 27 lines · Julia Lawall
Quoted 17 lines · Daniel Wagner
Quoted 4 lines · Julia Lawall
Quoted 4 lines
I am still struggling with creating a rule for a prototype. Rereading examples
and documentation.Quoted 8 lines
Quoted 4 lines
So I am only allowed to write a prototype (funproto, right?) in the declaration section (
common_decl)? How do I write the rule?(I am really bad not getting this done by myself.)
Could you send your attempt? It should be possible to write a function prototype pattern just like any other pattern, but it is true that I haven't done that very much, so it might be deficient in some way.
So my approach was to try this here:
Quoted 7 lines
@r@
identifier f;
parameter list[n] ps;
identifier i;
@@
f(ps, gboolean i, ...);
Quoted 8 lines
Just add the metavariable declaration
type T;
and T to the left of f in the prototype. For function definitions, it is OK with the return type being absent, but apparently not for prototypes. Leaving it out probably causes conflicts for the parser, because a prototype looks like a function call.
Yes, that did the trick. Thanks again!
cheers,
daniel